Since 2016, I’ve taken water samples from the outlets of tile lines draining fields across southeast Minnesota. The goal was to determine how much nitrate-nitrogen was leaching out of fields through drain tile. What started off as taking samples every two weeks from three field tiles in 2016 expanded to sampling 27 tiles biweekly by 2021. This project has yielded some interesting observations.
The sites
Most of the tile outlets (23 of 27) sampled drained fields that alternated between corn and soybeans every year. Of the other four sites, one had rotations that included corn/soybeans/grazed cool-season grasses/sweet corn, one (with just one year of sampling) was fallow after corn and soybeans, one was generally a multispecies hay mix (usually including clovers, alfalfa, grasses, and brassicas), and one rotated between corn, soybeans, and sweet corn. Only a few fields received manure applications during the sampling period. A couple other interesting aspects of the sampled sites were that a significant number of sites were strip- or no-tilled, and quite a few (nine of 23) of the corn/soybean fields were cover cropped for many or all of the studied years (spring 2016 through spring 2022).
Some observations
1. Sampled tiles often had nitrate concentrations that were lower than expected.
Figure 1 shows nitrate concentrations for the tiles sampled for this project. As a reference, the drinking water standard for nitrates is 10 milligrams per liter (mg/L). Above that level, water is not considered safe to drink, and isn’t allowed to be used as a public water supply without treatment. Tile systems in this study were more often than not below this standard, which is good news. In many ways, it was surprising, though, as many studies and sources of data (
Purdue University,
Discovery Farms MN,
Minnesota Pollution Control Agency,
Minnesota Department of Agriculture) would suggest that in a corn/soybean rotation (which is what most of the sampled sites were), nitrate concentrations of 15-20 mg/L is about average in tiles, although some other studies would peg that average slightly lower (
Iowa Soybean Association). There are several reasons that might help explain why nitrate concentrations were so low- the wide use of cover crops (which have been shown to greatly lower nitrate losses) and the high prevalence of minimum/no-tillage (which has been shown to lower nitrate
concentrations, if not total nitrate losses) come to mind. Another factor that likely comes into play is the high rainfall during the study time, which tends to lower the concentration of nitrates in drainage water. Lastly, there is some reason to suspect that not all the water coming into the tile system was coming from the drained field itself (more on this in observation #3 below)
 |
Figure 1. Nitrate concentrations for 27 southeast
Minnesota tile lines over 6 years.
Samples were collected ~biweekly, with exceptions during the winter of
2016, spring of 2020, and fall of 2021.
While the amount of data in this graph can obscure some trends, I will
note a few things. First, the majority
of samples collected had nitrate concentrations lower than 10 mg/L, the
drinking water standard. Second, there
was a general trend of decreasing nitrate concentrations in tile lines from
2016 to 2020 (with one major exception being the spring of 2019), particularly
at sites with higher overall nitrate concentrations. Nitrate concentrations appear to be rising
since 2020. Neither the increasing nor
decreasing trends appeared at every location.
Nitrate concentrations represented as 2 mg/L on this graph may be lower
than indicated, as that was the bottom end of detection for the method I used
to get nitrate results. |
2. Nitrate concentrations generally did not change rapidly over time.
They could fluctuate during and after rain events, and they tended to change somewhat across the growing season, but the main changes were across the years. In general, nitrate concentrations declined from 2016 to 2020, and appear to be increasing again since 2020. This is probably related to the high precipitation during 2016-2019 (which diluted nitrate concentrations in large volumes of tile flow) and the dry weather during the 2020 and 2021 growing seasons (which probably concentrated nitrates in the smaller volume of tile flow). These changes, however, were relatively minor. Based on flow measurements, overall nitrogen losses in tile were much higher in wet years than in dry, despite the lower concentrations. The most consistent seasonal changes in nitrate concentration were the following: A) an increase in nitrate concentrations during April and May and/or B) a gradual increase in nitrate concentration as flow rates declined over the growing season. Generally, these changes in concentration were small (a few mg/L). There were some intriguing exceptions to this general observation of small, gradual changes in concentration over time. The most striking was during the spring of 2019, when a number of tile drains I was sampling suddenly had dramatic spikes in nitrate concentrations during late May (see Figure 2). High nitrate losses through tile do not necessarily mean a field needed more nitrogen fertilizer than usual that year. I make this comment not based specifically on the data from this project, but from on-farm nitrogen fertilizer trials I was involved with during the same time period. As an example, 2016 was a very high nitrate loss year in southeast Minnesota, but nitrogen research trials around the region generally showed that corn fields had relatively low needs for nitrogen fertilizer that year. We attributed this largely to the timing of nitrogen losses. Excessive rains (and nitrate leaching) occurred in July, August, and September of that year, while the spring had been normal-to-dry. Research has shown that high nitrogen losses in the early-to-mid season can reduce crop yields most. When losses occur later in the summer, high rates of nitrogen mineralization can generally compensate for high nitrogen losses due to excessive rains.
 |
Figure 2. Nitrate concentrations during the spring of 2019 for a subset of tile sampling locations. All of these sites were close to one another (within 10 miles), and many exhibited a large rise in nitrate concentration during April and May (as much as 15 mg/L) that was otherwise never seen during the 6 year study period. I’ll note that 2019 was the wettest year on record in the area (just west of Rochester) where these tiles were located. Also of interest is that, while all but one of these sites (the orange X’s on the graph) were corn/soybean rotation, many normally had cover crops. The fall of 2018 was very wet and cold, with a delayed harvest that prevented most from being seeded with covers, and minimized cover crop growth in those that were seeded. Of the two locations that did not have large spikes in nitrate concentrations (black “pluses” and orange “X’s”), the black pluses were from a field (discussed later) that is likely draining lots of water from outside the field boundary, and the orange X’s were from a field that had September-harvested sweet corn and a well-established cover crop grown after it. Dotted lines connecting the sampling points are merely for visual clarity and do not represent any statistical trend analysis. |
|
3. Flow rates and duration of tile flow can vary widely from drainage system to drainage system, even when tiled areas are similar in size.
Around 20% of the sites I sampled never or rarely stopped flowing, even during the 2021 drought. And while most sites showed a “typical” tile drainage pattern of flowing heavily during the spring and either tapering off or drying up during the summer, several sites flowed only sporadically- briefly during the spring or after significant rainfall events. In some cases tile systems appeared to be picking up water from outside of their drained area. In the most extreme example from this project, I took flow measurements for a specific 26-acre field for five years. This tile never stopped flowing during this time, and I measured an average flow of 1.73 gallons/second. If you accept that as its average flow rate during the five years, this tile drained a whopping 77.5 inches of water per year on average! That amounts to just over double the total precipitation average for this period of time. This field is directly downstream of a flood control reservoir, and is at the foot of some large upland areas, which leads me to believe groundwater from other sources is getting into the tile system.
 |
Picture 1. Water continuing to flow out of a tile main on the second-coldest day of 2019. |
4. There were greater winter flow rates than expected.
As I mentioned before, there were wide differences in the amount of water flowing through different tile systems. One thing that surprised me was how much winter tile flow I observed. Generally, any tiles that were still flowing in late summer and autumn continued to flow all winter long, regardless of how cold the weather was or how deep the frost layer. One of the more extreme examples comes from the winter of 2019. In late January, after a two-week period where average air temperatures were about -1˚F, we had two days of deep freeze, where temperatures got down to near -40 ˚F. On the second morning, I went out and checked several tiles. They were still flowing! (see Picture 1). Clearly, at least in southeast Minnesota, the paths water takes to get to these tile systems is somewhat shielded from weather conditions. The 2019 cold snap followed a wet fall, but a similar pattern of continual tile flow was observed during the coldest days of winter 2022, which followed a drought (the tiles that were flowing before the ground froze continued to flow all winter). Figure 3 shows the breakdown of average flow rates from one tile during non-frozen conditions versus frozen ground times of year (these were approximated based on regional soil temperature data). While flow rates were higher during non-frozen seasons, the difference was less than I would have expected.
 |
Figure 3. Tile flow at different times of year for one
tile main, seen during the winter in Picture 1.
The blue box and whisker shows the average, median, and range of tile
flow during times of year when the ground was frozen during a 5 year period,
while the orange box and whisker shows the same information during unfrozen
conditions. While overall flows were
higher during unfrozen conditions, the differences were less than expected
between these two categories. It’s worth
noting that there were times where I could not obtain flow rate measurements
(either because of too much ice accumulation during winter, or water being too
high in the ditch). It’s also worth
noting that flows above ~5 gallon/second were less accurate than slower flows
using my high-tech method of using a 5 gallon bucket and stopwatch to measure
flow rate. Lastly, nitrate
concentrations (not shown) were higher during frozen soil conditions, which
further reduces the difference in overall nitrate losses between frozen and
unfrozen conditions (since loss= flow times concentration). |
5. Farm ponds can potentially remove a lot of nitrate from a tile system.
There was one farm I started taking limited samples from in July 2021 where two tiles entered a small (~1/10th acre) farm pond, which then flowed into a small creek (see Picture 2). I took water samples from the two tiles, as well as the pond outlet. During the height of summer, the flow into and out of the pond was slower, the water was warm, and there was a lot of vegetation in the pond, As seen in Figure 4, nitrate concentrations at the pond outlet were substantially lower (~ 70%) than in the tiles flowing into the pond during this time. This suggests that nitrate was being removed at high rates by the plants or microorganisms in the pond. This nitrate removal basically disappeared by late fall, when nitrate concentrations at the pond outlet were similar to concentrations at the tile outlets.
 |
Picture 2. A small (~1/10th acre) farm pond that receives water from two tile systems (out of sight below the left-hand corner of the picture), and discharges water into a small stream on the far side of the pond (barely visible at the top of the picture as an area of rip-rap). This pond, despite its small size, had significant impacts on nitrate concentrations during ideal conditions for nitrate uptake/reduction.
|
 |
Figure 4. Nitrate
concentrations in two tile outlets (orange and blue points/lines) that drain
into a farm pond, as well as nitrate concentrations in the pond outlet (gray
squares). For most of summer 2021, the
reduction in nitrates at the pond outlet vs. the tile outlets was substantial,
but during the fall, nitrate concentrations in the pond outlet appeared to revert
to a largely unchanged mix of the tile water sources. During spring 2022, there hasn’t appeared to
be much nitrate reduction, and overall nitrate concentrations are rising
substantially. |
6. You can get fairly high-quality information, cheaply, about your average nitrate losses from these types of water samples.
Research from Discovery Farms Minnesota indicates that, for nitrate, sampling a tile every 2 weeks gives you similarly high quality data to expensive automated machines that conduct continuous flow monitoring. I would even go so far as to say that if you can grab a sample every month during the growing season, it can tell you a lot about your average field nitrate losses. There are also many places you can take water samples to get analyzed for nitrates. Many Soil and Water Conservation Districts have access to equipment for analyzing nitrates in water samples; and the Minnesota Agricultural Water Resources Center and other farm organizations occasionally put on events where they do this sort of analysis for farmers. If you prefer, there are also many relatively cheap options you can purchase for your own use (from test strips to handheld devices) that may give you good estimates of your tile’s nitrate concentrations. While weather conditions and questions about where the water in a tile system is originating can complicate the interpretation of the nitrate data you get, they can help serve as a valuable benchmark. And you might be pleasantly surprised by what you find.
A sincere “thank you” to every farmer and landowner who participated in this project, as well as the Southeast Minnesota Water Analysis Laboratory and Fillmore Soil and Water Conservation District for access to equipment to conduct water analysis on. Lastly, thank you to the Minnesota Department of Agriculture for providing the sample analysis equipment to the labs.
Comments
Post a Comment